Fig. 2: Anti-VEGF treatment inhibits proliferation, but not inflammatory responses or induction of GSP58 signature, in metastasis-associated ECs.
a, Experimental outline of anti-VEGF treatment of mice with metastatic lung nodules. MDA231-LM2 cells were injected intravenously into NSG mice followed by repeated treatment with either anti-VEGF antibody B20.4.1.1 (B20) or control IgG for 3 weeks. Lung ECs were isolated at week 3 and transcriptomic analysis was performed. Mice harboring comparable lung metastatic loads, as measured by in vivo bioluminescence imaging, were selected for analysis. b, Overview of cell function signatures (REACTOME) repressed in metastasis-associated ECs by anti-VEGF treatment. Signatures with FDR < 0.05 were included in calculations of percentages. c, GSEA of gene clusters: cell cycle mitosis (REACTOME), DNA replication (REACTOME), sprouting angiogenesis (C5 collection in MSigDB) and tip cell signature12 in lung ECs after B20 treatment compared with control IgG. FDR was determined from P values calculated by random permutation test. d,e, Violin plots showing z-scores of genes from a lung EC proliferation signature associated with metastasis (Supplementary Table 3) (d) and tumor angiogenesis UP cluster16 (e) expressed in purified lung ECs from control (ctrl) or metastasis (met.)-bearing mice with the indicated treatment. f, Downregulated gene signatures in metastasis-associated ECs treated with B20 compared to IgG control, based on GSEA. Signatures with FDR < 0.25 are shown. Blue, proliferation-related signatures; red, signatures of VEGF target genes; gray, others. g–i, Violin plot analysis of stroma poor-outcome cluster13 (g, left) and angiogenesis genes associated with poor prognosis14 (g, right); lung EC inflammation signature associated with metastasis (Supplementary Table 4) (h) and GSP58 (i) expressed in ECs from mouse lungs under the indicated conditions. In d,e and g–i, P values were determined from averaged z-score of genes within signatures by one-way ANOVA with Dunnett’s multiple comparison test; n = 3 mice per group. NS, not significant.
